CY gate (controlled-Y) is a two-qubit gate that applies a Pauli Y to the target qubit if and only if the control qubit is $\lvert 1\rangle$, and does nothing otherwise.
$$\text{CY} = \begin{pmatrix}1 & 0 & 0 & 0\\ 0 & 1 & 0 & 0\\ 0 & 0 & 0 & -i\\ 0 & 0 & i & 0\end{pmatrix}$$
The rows and columns are ordered $\lvert 00\rangle, \lvert 01\rangle, \lvert 10\rangle, \lvert 11\rangle$, with the first qubit as control and the second as target:
$$\text{CY}\lvert 00\rangle = \lvert 00\rangle \qquad \text{CY}\lvert 01\rangle = \lvert 01\rangle \qquad \text{CY}\lvert 10\rangle = i\lvert 11\rangle \qquad \text{CY}\lvert 11\rangle = -i\lvert 10\rangle$$
A Hadamard gate on the control followed by CY produces an entangled state from $\lvert 00\rangle$. The resulting state differs from the Bell states by a local phase on the target: $\frac{1}{\sqrt{2}}(\lvert 00\rangle + i\lvert 11\rangle)$. Like the CX gate, CY is an entangling gate — the output cannot be written as a product of two single-qubit states.
The CY gate is its own inverse up to global phase: $\text{CY}^2 = I$. It is equivalent to the CX gate up to single-qubit rotations on the target: $\text{CY} = (I \otimes S) \cdot \text{CX} \cdot (I \otimes S^\dagger)$, where $S$ is the S gate. Together with all single-qubit gates it forms a universal gate set for quantum computation.